2D混合MIMO相控阵雷达收发阵列结构的交错稀疏优化设计理论与方法

Although conventional coherent MIMO radars with collocated antennas offer unique advantages over their phased-array counterparts through waveform diversity gain stemming from orthogonal waveforms transmission, the system cost paid for these potential performance claims is increased prohibitively for the two-dimensional (2D) array geometry with large number of transmit elements and concurrently the remarkable loss of coherent transmit processing gain due to omnidirectional transmission has considerably limited their practical applications in some radar operation modes. Lower hardware complexity and a beneficial performance tradeoff between waveform diversity gain and coherent transmit processing gain can be achieved with Hybrid MIMO phased-array radar (HMPAR) using transmit subarray splitting and weighting. Due to its lower system cost and larger subaperture size, interleaved sparse elements arrangement method is proposed in this project for the array geometry design of HMPAR with a view to acquire the optimal performance tradeoff between the size of effective virtual aperture and SNR (signal-to-noise) on receive array. In light of the fact that the transmit subarray splitting in the HMPAR can be equivalently regarded as a special design of transmit waveform covariance matrix, a bridge will be established in this project using transmit waveform covariance matrix between the array geometry design and system performance analysis. Further the mathematical model of array steering vector of virtual receive array and its manifold analysis underlie the theoretical framework of system performance analysis for HMPAR with interleaved sparse array geometry. The equivalent model of multiple orthogonal and disjoint tasks learning in the powerful mathematical tools of Multi-Task Learning (MTL) is employed in our project to effectively interleave several sparse subarrays with favorable aperture size and beampatterns. The Bayesian Compressive Sensing (BCS) and Matrix Pencil Method (MPM) are exploited to realize the two-way objective beampattern matching of HMPAR. The coordinated transmit-receive beamforming and the 2D high-resolution direction-of-arrival (DOA) estimation based on the steering vector of resultant virtual receive array are explored in details to evaluate the overall and compromised performance between waveform diversity gain and coherent transmit processing gain acquired with our proposed interleaved sparse array geometry.

2D相干MIMO雷达的实现代价高，波形分集增益与发射相干处理增益之间的矛盾尤为突出。本项目利用交错稀疏阵元设置方法在实现代价和孔径利用率方面的优势，将交错稀疏阵列设计方法引入2D混合MIMO相控阵雷达收发阵列结构及其发射子阵分割的设计中。基于发射阵列结构与发射波形协方差矩阵的等效关系，拟通过阵列结构优化实现2D相干MIMO雷达波形分集增益和发射相干增益的最佳折中，有效解决接收端信噪比和虚拟阵列孔径扩展之间的矛盾。项目的研究以等效发射波形协方差矩阵的设计为桥梁，以接收虚拟阵列导向矢量的建模、分析与应用为主线，构建收发阵列结构对2D混合MIMO相控阵体制性能影响分析的理论框架；利用多任务学习中的“多正交独立任务学习”模型和方法实现2D混合MIMO相控阵雷达收发阵列结构的稀疏交错优化设计；通过收发联合波束赋形和基于虚拟阵列导向矢量的2D超分辨波达方向估计性能来验证和评估阵列结构优化设计的有效性。

混合MIMO相控阵体制可以在有效简化系统实现的同时实现雷达波形分集增益和发射相干增益的有效折中。交错稀疏阵列形式是混合MIMO相控阵体制理想的阵列结构。混合MIMO相控阵体制收发阵列的交错稀疏阵列设计面临的主要难题是收发结构设计与发射子阵分割的分级设计。本项目围绕2D混合MIMO相控阵雷达收发阵列结构的交错稀疏优化设计理论与方法，在构建起2D混合MIMO相控阵雷达交错稀疏阵列设计的优化准则的基础上，对2D混合MIMO相控阵雷达的交错稀疏阵列结构进行优化设计，最后对交错稀疏的2D混合MIMO相控阵雷达的性能进行全面评估。具体地，利用压缩感知、多任务学习、嵌套阵、卷积神经网络等理论和方法，针对发射阵列稀疏优化设计、交错稀疏模型与方法、混合MIMO相控阵雷达发射子阵分割、混合MIMO相控阵雷达阵列结构优化设计和面向2D混合MIMO相控阵雷达的空间信源超分辨波达方向估计进行了较深入的研究。取得的主要研究成果包括：（1）面向发射方向图赋形采用酉矩阵束、贝叶斯压缩感知、多任务学习等方法进行了发射阵列的稀疏优化设计，为2D混合MIMO相控阵雷达交错稀疏阵列结构优化设计打下基础；（2）基于区域约束贝叶斯压缩感知和频谱能量分配实现了阵列交错稀疏优化布阵；（3）利用卷积神经网络、中心扩展、反向非均匀等理论和方法对2D混合MIMO相控阵雷达的发射阵列子阵分割进行优化设计；（4）聚焦混合MIMO相控阵雷达的收发阵列结构优化设计，利用嵌套阵、十字形阵等特殊阵列结构实现了混合MIMO相控阵雷达的收发阵列结构联合优化设计；（5）研究了一系列面向2D混合MIMO相控阵雷达的波达方向估计算法，为2D混合MIMO相控阵雷达整体性能评估与实用提供理论与技术基础。课题取得的一系列有特色、较有影响的研究成果为我国实用化的2D混合MIMO相控阵雷达的交错稀疏阵列结构设计及其应用奠定了相关理论基础。项目按照研究计划进行，取得的研究成果主要反映在已撰写的23篇学术论文中，目前已发表和录用待发表的18篇学术论文中，SCI源刊论文6篇，EI源刊论文12篇。